AMDGPU: Remove custom getSubReg

[arsenm]

Summary

This was kind of confusing, the subregister class shouldn't really be necessary.

[nhaehnle]

LGTM

[arsenm]

r278362

lib/Target/AMDGPU/SIRegisterInfo.h

public:
   const TargetRegisterClass *getSubRegClass(const TargetRegisterClass *RC,
                                             unsigned SubIdx) const;
 
   bool shouldRewriteCopySrc(const TargetRegisterClass *DefRC,
                             unsigned DefSubReg,
                             const TargetRegisterClass *SrcRC,
                             unsigned SrcSubReg) const override;
 
-  /// \p Channel This is the register channel (e.g. a value from 0-16), not the
-  ///            SubReg index.
-  /// \returns The sub-register of Reg that is in Channel.
-  unsigned getPhysRegSubReg(unsigned Reg, const TargetRegisterClass *SubRC,
-                            unsigned Channel) const;
-
   /// \returns True if operands defined with this operand type can accept
   /// a literal constant (i.e. any 32-bit immediate).
   bool opCanUseLiteralConstant(unsigned OpType) const;
 
   /// \returns True if operands defined with this operand type can accept
   /// an inline constant. i.e. An integer value in the range (-16, 64) or
   /// -4.0f, -2.0f, -1.0f, -0.5f, 0.0f, 0.5f, 1.0f, 2.0f, 4.0f.
   bool opCanUseInlineConstant(unsigned OpType) const;

lib/Target/AMDGPU/SIRegisterInfo.cpp

if (!isUInt<12>(Offset + Size)) {
     }
     BuildMI(*MBB, MI, DL, TII->get(AMDGPU::S_ADD_U32), SOffset)
             .addReg(ScratchOffset)
             .addImm(Offset);
     Offset = 0;
   }
 
   for (unsigned i = 0, e = NumSubRegs; i != e; ++i, Offset += 4) {
-    unsigned SubReg = NumSubRegs > 1 ?
-        getPhysRegSubReg(Value, &AMDGPU::VGPR_32RegClass, i) :
-        Value;
+    unsigned SubReg = NumSubRegs == 1 ?
+      Value : getSubReg(Value, getSubRegFromChannel(i));
 
     unsigned SOffsetRegState = 0;
     unsigned SrcDstRegState = getDefRegState(!IsStore);
     if (i + 1 == e) {
       SOffsetRegState |= getKillRegState(Scavenged);
       // The last implicit use carries the "Kill" flag.
       SrcDstRegState |= getKillRegState(IsKill);
     }
case AMDGPU::SI_SPILL_S32_SAVE: {
       unsigned NumSubRegs = getNumSubRegsForSpillOp(MI->getOpcode());
       unsigned TmpReg = MRI.createVirtualRegister(&AMDGPU::VGPR_32RegClass);
 
       unsigned SuperReg = MI->getOperand(0).getReg();
       bool IsKill = MI->getOperand(0).isKill();
       // SubReg carries the "Kill" flag when SubReg == SuperReg.
       unsigned SubKillState = getKillRegState((NumSubRegs == 1) && IsKill);
       for (unsigned i = 0, e = NumSubRegs; i < e; ++i) {
-        unsigned SubReg = getPhysRegSubReg(SuperReg,
-                                           &AMDGPU::SGPR_32RegClass, i);
+        unsigned SubReg = NumSubRegs == 1 ?
+          SuperReg : getSubReg(SuperReg, getSubRegFromChannel(i));
 
         struct SIMachineFunctionInfo::SpilledReg Spill =
             MFI->getSpilledReg(MF, Index, i);
 
         if (Spill.hasReg()) {
           BuildMI(*MBB, MI, DL,
                   TII->getMCOpcodeFromPseudo(AMDGPU::V_WRITELANE_B32),
                   Spill.VGPR)
switch (MI->getOpcode()) {
     // SGPR register restore
     case AMDGPU::SI_SPILL_S512_RESTORE:
     case AMDGPU::SI_SPILL_S256_RESTORE:
     case AMDGPU::SI_SPILL_S128_RESTORE:
     case AMDGPU::SI_SPILL_S64_RESTORE:
     case AMDGPU::SI_SPILL_S32_RESTORE: {
       unsigned NumSubRegs = getNumSubRegsForSpillOp(MI->getOpcode());
       unsigned TmpReg = MRI.createVirtualRegister(&AMDGPU::VGPR_32RegClass);
+      unsigned SuperReg = MI->getOperand(0).getReg();
 
       for (unsigned i = 0, e = NumSubRegs; i < e; ++i) {
-        unsigned SubReg = getPhysRegSubReg(MI->getOperand(0).getReg(),
-                                           &AMDGPU::SGPR_32RegClass, i);
-        struct SIMachineFunctionInfo::SpilledReg Spill =
-            MFI->getSpilledReg(MF, Index, i);
+        unsigned SubReg = NumSubRegs == 1 ?
+          SuperReg : getSubReg(SuperReg, getSubRegFromChannel(i));
+
+        SIMachineFunctionInfo::SpilledReg Spill
+          = MFI->getSpilledReg(MF, Index, i);
 
         if (Spill.hasReg()) {
           BuildMI(*MBB, MI, DL,
                   TII->getMCOpcodeFromPseudo(AMDGPU::V_READLANE_B32),
                   SubReg)
                   .addReg(Spill.VGPR)
                   .addImm(Spill.Lane)
                   .addReg(MI->getOperand(0).getReg(), RegState::ImplicitDefine);
bool SIRegisterInfo::shouldRewriteCopySrc(
   //
   // We want to look through the COPY to find:
   //  => vreg3 = COPY vreg0
 
   // Plain copy.
   return getCommonSubClass(DefRC, SrcRC) != nullptr;
 }
 
-unsigned SIRegisterInfo::getPhysRegSubReg(unsigned Reg,
-                                          const TargetRegisterClass *SubRC,
-                                          unsigned Channel) const {
-
-  switch (Reg) {
-    case AMDGPU::VCC:
-      switch(Channel) {
-        case 0: return AMDGPU::VCC_LO;
-        case 1: return AMDGPU::VCC_HI;
-        default: llvm_unreachable("Invalid SubIdx for VCC"); break;
-      }
-
-    case AMDGPU::TBA:
-      switch(Channel) {
-        case 0: return AMDGPU::TBA_LO;
-        case 1: return AMDGPU::TBA_HI;
-        default: llvm_unreachable("Invalid SubIdx for TBA"); break;
-      }
-
-    case AMDGPU::TMA:
-      switch(Channel) {
-        case 0: return AMDGPU::TMA_LO;
-        case 1: return AMDGPU::TMA_HI;
-        default: llvm_unreachable("Invalid SubIdx for TMA"); break;
-      }
-
-  case AMDGPU::FLAT_SCR:
-    switch (Channel) {
-    case 0:
-      return AMDGPU::FLAT_SCR_LO;
-    case 1:
-      return AMDGPU::FLAT_SCR_HI;
-    default:
-      llvm_unreachable("Invalid SubIdx for FLAT_SCR");
-    }
-    break;
-
-  case AMDGPU::EXEC:
-    switch (Channel) {
-    case 0:
-      return AMDGPU::EXEC_LO;
-    case 1:
-      return AMDGPU::EXEC_HI;
-    default:
-      llvm_unreachable("Invalid SubIdx for EXEC");
-    }
-    break;
-  }
-
-  const TargetRegisterClass *RC = getPhysRegClass(Reg);
-  // 32-bit registers don't have sub-registers, so we can just return the
-  // Reg.  We need to have this check here, because the calculation below
-  // using getHWRegIndex() will fail with special 32-bit registers like
-  // VCC_LO, VCC_HI, EXEC_LO, EXEC_HI and M0.
-  if (RC->getSize() == 4) {
-    assert(Channel == 0);
-    return Reg;
-  }
-
-  unsigned Index = getHWRegIndex(Reg);
-  return SubRC->getRegister(Index + Channel);
-}
-
 bool SIRegisterInfo::opCanUseLiteralConstant(unsigned OpType) const {
   return OpType == AMDGPU::OPERAND_REG_IMM32;
 }
 
 bool SIRegisterInfo::opCanUseInlineConstant(unsigned OpType) const {
   if (opCanUseLiteralConstant(OpType))
     return true;